Project Background
Tunnel Unterwald is a double track railway tunnel with a length of 1076m and a standard cross section of 100m2. The excavation was carried out according to the NATM with round lengths between 1 and 2.2 m. The tunnel was parallel to the hill slope with an overburden up to 90m. It is located in central Austria and is part of the railroad section Schoberpass, which has the lowest elevation crossing of the Alps (840 m). This railway line from Amstetten to Tarvisio (Italy), connecting Northern and Southern Austria, is a very important transit route and is currently being upgraded to a high-capacity railway.A field reconnaissance investigation was performed in 1994 and 1995 followed by a detailed site investigation programme. Subsurface investigations were performed from 1994 to 1996. Twenty six cored vertical, inclined, and horizontal boreholes with a total length of 1 600 m were drilled. In situ- and laboratory tests were carried out in order to gain reliable rock and soil parameters. To supplement the subsurface exploration geophysical investigations including refraction seismic, electrical resistivity, electromagnetic surveys were performed.
Geology
The project area is situated within the Rannach Series, a metamorphic rock unit of Permo-mesozoic age characterized by highly anisotropic rocks with a varying quartz content. The rock mass primarily consists of foliated quartzite and quartz-phyllite with occasional intercalations of sericitic phyllite, chloritic phyllite and carbonaceous phyllite. The foliation strikes parallel to the tunnel and dips on average 25° to 35°, which is parallel to the hill-slope. The micaeous content is crenulated and larger scale folding of the foliation occurs with a similar fold axis as the crenulation, parallel to the foliation strike. Occasionally foliation parallel shear zones were encountered. Two dominant joint sets were encountered striking approximately perpendicular to the tunnel axis and dipping steeply either into or away from the tunnel face. A few minor faults were encountered which were parallel to the dominate joint sets.
The portal areas are located within young sediments consisting of glacial-fluvial gravel and debris with large boulders in the East and landslide debris consisting of boulders and large rock slabs separated by silty-gravelly shears in the West.
TRT imaging results
During the initial site investigation, the first 170 to 200 m of the excavation were predicted to be in slope debris consisting of a dense blocky soil and the exact boundary to the competent bedrock was not specifically located. In order to appropriately plan the excavation methods and to have the appropriate support stocks on site knowing this boundary precisely would save considerable costs.
This was the first time that the TRT method was applied in soft extremely heterogeneous ground conditions with weak seismic contrasts and in highly anisotropic material. It was determined that the maximum reliable imaging distance was between 60 and 70 m (compared to 100 to 150 m expected in hard crystalline rock) due to the high attenuation associated with the blocky material and the soft matrix.
The second survey was planned so that the imaged distance would include the initially predicted region for the bedrock contact. The survey was performed when the tunnel excavation was at tunnel station 152 m, resulting in an image to approximately station 210 to 220 m.
Initially, a dual velocity model was used with a high velocity bedrock below the tunnel but the data was inconclusive. Figure 1 identified a large boulder approximately 5 m ahead of the tunnel face. Figure 2 shows a photo taken of the tunnel face at station 161 m which corresponds to the image results.
Fig. 1 Plan view of horizontal TRT tomogram at 152m Fig.2 Large boulder of quartz-phyllite encountered at station 161m
Fig.3 Bedrock contact at station 191-192m
The results also indicated an positive
anomaly (to a higher velocity) at approximately station 190 m to the right of
the tunnel axis. This anomaly was initially interpreted to be representative of
the bedrock contact but was connected to the series of smaller anomalies
forming a contact at an acute angle to the tunnel axis. It turned out that the
smaller anomalies were individual features and the large anomaly was the
bedrock contact which was orientated approximately perpendicular to the tunnel
axis as schematically drawn on the tomogram. Figure 3 shows the encountered
bedrock contact at station 191 to 192 m. The contact was formed by discontinuities
of large planar wall like joints. The bedrock was slightly weathered with
widely spaced (approximately 1 m) open joints striking almost perpendicular to
the tunnel face and dipping steeply to the East. The joints were filled with
sandy-silty material with a thickness up to 20 cm.
Imaging a fault zone and image repeatability
The seventh and eighth surveys serve as an
example of the repeatability and the potential accuracy of the TRT method. The
results of the seventh survey indicated a fractured zone approximately 25 m
ahead of the excavation followed by relatively homogeneous ground conditions.
Figure 4 shows the results of a second processing using a different attenuation
model to look further ahead of the excavation.
Fig. 4 Fig. 5
It shows an anomaly around stations 825 to 840 m, almost 150 m from the tunnel face. The initial site investigation had provided evidence that there was a probable fault zone in this region. The anomalies imaged during the seventh survey compared quite well with encountered geological conditions. The eighth survey was performed with the face at station 771 m, the initial velocity model did not indicate any anomalies ahead of the face. Therefore, a second velocity and attenuation model was used and several anomalies were identified from stations 785 to 835 m as shown in Figure 5.
The first anomalies were described as a possible fracture zone while the anomalies from station 820 to 840 m were interpreted to be a fault zone. The encountered conditions consisted of two sets of sub-vertical discontinuities spaced between 0.5 to 1 m crossing the tunnel at an angle of 60° to 70°. At station 820 m the tunnel entered a 8 m wide sharply defined shear-zone, steeply dipping 70° to the East and striking perpendicular tunnel face. This fault consisted of lens like anastomosing shear-bodies formed by two sets of slickensided planes. The location of this fault corresponded to the feature imaged during the seventh survey. The difference in the amount of information between the two surveys shows how different attenuation models can be used to recognize either smaller features closer to the excavation or larger features at much greater distances. The anomalies identified in this survey and their interpretations correlated very well with the encountered geological conditions as shown in the previous figures.
Conclusion
As shown with the two examples above the TRT method can provide accurate predictions of changes in the rock mass conditions. It is in our opinion that more work needs to be incorporated into the reliability and use of the TRT data to appropriately predict what conditions will be encountered during the excavation and if the changing ground conditions actually effect the excavation behavior. This also would include more information on the extrapolation of features not directly in the path of the tunnel to an appropriate position.
This was the first time that the TRT method was used in the ground conditions encountered during the first 200 m of the excavation and some initial difficulties had to be overcome. The soft and blocky nature of the soil mass resulted in a very rapid attenuation of the seismic signal considerably reducing the reliable image lengths. . A second problem was delivering enough energy into the ground to compensate for the increased attenuation. Because the soil was “softer” then the lining or the large blocks the energy delivered by the source was not efficiently transferred into the soil also resulting in decreased image lengths.
This was also the first time that the TRT method was used in an highly anisotropic material. Since the strike of the foliation was parallel with the survey direction the anisotropic nature of the material was generally avoided but may have contributed to the localized imaging of features that were quite extensive compared to the survey dimensions. The rock mass was rather homogeneous except for the portal locations. With the low overburden and relatively good rock mass conditions deformations over a majority of the tunnel were minimal and did not change significantly in the predicted zones of weaker material. Instead, the fault zones contributed to increased overbreak and occasionally to increased water flow. The gradual increase in fracture density that proceeded the fault zones generally contributed to the smoothing of the reflection data making features less pronounced.
The True Reflection Tomography seismic imaging method was used at the Unterwald tunnel on a systematic basis to identify potential weakness zones intersecting the tunnel axis. Generally, the data was processed several times to clarify the results and uncertainties in the data interpretations. As with any seismic method, one must use all of the available information to interpret the results. Sometimes this philosophy can be misleading if too much is read into the data from pre-existing models as demonstrated with the example showing the bedrock contact. When integrating this method into the tunnel construction the reliability of the data should be quantified so the engineer can feel confident with the interpretation of the results. This requires communication and cooperation between the data processors, the geologists, and the engineer to discuss not only the current data results, but also the past evaluations so improved forward predictions can be made for the given the site specific conditions.
Overall, the TRT method provided good results as most of the major features were imaged. The confirmation of the location of features that are generally, but not specifically known, is possible with this method as well as locating unexpected geological conditions, but it must be stressed that proper communication must exist between all the parties involved to delineate the strengths and weakness of the system in a given geologic environment. This is still a relatively young technology that will improve with time as more experience is gained in different ground conditions.